Field
The present invention relates to a device and method for recognizing an obstacle around a vehicle, and more particularly, to a device and method capable of improving performance of recognizing an obstacle around a vehicle.
Discussion of the Background
In general, for safe driving of an autonomous vehicle, it is very important to detect an obstacle in a driving direction and not to collide against the obstacle based on information collected regarding the detected obstacle.
For example, when driving on the road, the autonomous vehicle should drive in a manner so that it does not collide against a sidewalk protruding higher than the road. Thus, detecting a curb which is a boundary stone between the road and the sidewalk is needed. The curb is an important sign on the road. Thus, accurate detection of a curb is useful in the field of autonomous vehicles.
In a method of detecting a curb according to the related art, the curb is detected mainly using sensor information obtained from side surfaces thereof. That is, in the related art, the curb is detected using a feature of distance sensor information which sharply changes at portions of the curb since the side surfaces of the curb are perpendicular to the ground.
However, the amount of the sensor information obtained from the side surfaces of the curb is relatively small. Thus, when only the sensor information obtained from the side surfaces of the curb is used, the performance of detecting the curb is likely to be influenced by external causes, sensor noise, or the like. Furthermore, it is difficult to recognize a curb distant from a vehicle. For example, side surfaces of a curb may be short, e.g., in the range of 6 cm to 15 cm, and thus a relatively small amount of information may be obtained from the side surfaces. Thus, when a distance to a curb is large, an amount of information enough to differentiate between the ground and the curb is difficult to acquire. Accordingly, detecting a curb distant from a vehicle is difficult according to the related art.
In another example, for safe driving, the autonomous vehicle detects an obstacle using signal information measured by a distance measurement sensor. In detail, the autonomous vehicle detects an obstacle by separating information regarding the ground on the road and information regarding an obstacle on the road from the measured signal information and extracting only three-dimensional (3D) information, i.e., the information regarding the obstacle, by removing distance information estimated as the information regarding the ground.
However, since the information regarding the obstacle is extracted by removing all the distance information estimated as information regarding the ground (low-level height information) from the signal information measured by the distance measurement sensor, it is likely that a curb having a low height and the like will be considered the ground, and thus information thereof will be removed. Furthermore, height information regarding an obstacle distant from the vehicle may be measured as having a low level and may be thus removed together with the information regarding the ground.
The above information disclosed in this Background section is only for enhancement of understanding of the background of the inventive concept, and, therefore, it may contain information that does not form the prior art that is already known in this country to a person of ordinary skill in the art.